UOB GCD GEO 101
Contents
Executive Summary 01 What Is GEO? 02 GEO vs SEO 03 How AI Engines Work 04 Machine-Readable Content 05 Research & Ideation 06 Enhancing Product Pages 07 All GEO Elements 08 Quick-Start Checklist A1 AEO Explained A2 E-E-A-T Deep Dive
📡 Internal Reference Guide · 2026

GEO 101: Optimising for the AI Era

A comprehensive guide to Generative Engine Optimisation.

8
Core Topics
15+
GEO Elements Explained
2
Appendix Deep Dives
Executive Summary

The way customers find financial products is changing — and we need to move with it.

Search is no longer just a list of links. A growing share of customers now ask AI tools — ChatGPT, Google AI Overviews, Microsoft CoPilot — for financial advice, product comparisons and recommendations. These AI engines generate direct answers, and they cite specific sources. If we are not among those sources, we are invisible at the most critical point in the customer's decision journey.

3–6
months for content optimisations to show up in AI citation behaviour
20%
of content changes deliver 80% of GEO improvement — the gains are concentrated
12
measurable checks determine whether a page is ready to be cited by AI engines
The Opportunity

AI search is already shaping customer decisions in our markets. Early movers who structure their content for AI citation will hold a compound advantage that late adopters will find difficult to close.

What GEO Is

Generative Engine Optimisation (GEO) is the discipline of structuring content so AI tools choose to cite us. It builds on our existing SEO investment — it is not a replacement, but a strategic extension.

What We Are Doing

We are tracking our AI citation performance using BrightEdge and restructuring priority pages for GEO readiness across business units.

What We Need

Cross-functional alignment to prioritise GEO content improvements and integrate GEO metrics into our digital KPI framework.

01

What Is GEO?

GEO stands for Generative Engine Optimisation. It is the practice of structuring and writing content so that AI-powered tools — like ChatGPT, Google AI Overviews, Microsoft CoPilot, and Google Gemini — choose to include our content in the answers they generate.

💡
The Shift That Changes Everything

SEO puts our website on Google's list of results. GEO puts UOB's name and knowledge inside the AI's answer — so customers are already considering us before they click anything.

The Problem GEO Solves

Imagine a potential UOB business customer who types into ChatGPT: "What's the best SME working capital loan?"

ChatGPT doesn't show a list of blue links. It generates a confident, paragraph-form answer — citing two or three banks by name, explaining their features, and sometimes even recommending one. That answer is formed in seconds. The customer reads it, shortlists those banks, and may never run a traditional Google search at all.

If UOB isn't in that answer — if we aren't cited — we have lost a customer before the funnel even began. GEO is how we make sure UOB is in the answer.

🌍 Where GEO Happens

  • Google AI Overviews (highest traffic)
  • ChatGPT (GPT-4o browsing mode)
  • Microsoft CoPilot
  • Google Gemini
  • Claude (Anthropic)

🎯 Why It Matters for Banks

  • Customers research loans, cards & savings via AI
  • AI citations build trust & brand familiarity
  • Citation links drive high-intent web traffic
  • Competitors actively optimising for AI search
  • Accuracy risk: AI may cite wrong rates or terms
  • No GEO = invisible in the AI research phase
⚠️
The Risk of Doing Nothing

AI search is not coming — it is already here. Google's AI Overviews appear for millions of queries daily across markets where UOB operates. Industry peers are actively producing content structured for AI citation. Every month without a GEO strategy is a month where others build a citation advantage that compounds over time.

02

GEO vs SEO — Understanding the Difference

GEO and SEO are complementary disciplines, not competing ones. Understanding how they differ — and where they overlap — is essential to building a content strategy that performs across both traditional and AI-powered search.

Traditional
SEO
Search Engine Optimisation
🎯Goal: Rank high on Google's blue-link results page
👤User intent: "Find me the best webpages"
📊Metric: Keyword ranking, organic clicks, impressions
🛠Tactics: Keywords, backlinks, page speed, meta tags
📈Attribution: Direct click tracking via GA4/Adobe Analytics
Speed: 3–6 months to see significant ranking change
⭐ New Priority
GEO
Generative Engine Optimisation
🎯Goal: Be cited inside AI-generated answers
👤User intent: "Answer, research, recommend for me"
📊Metric: Citation rate, mention rate, Share of Voice
🛠Tactics: FAQ structure, schema markup, authoritative answers
📈Attribution: Directional (no direct conversion tracking yet)
Speed: 3–6 months; some improvements visible faster
Built on the Same Foundation

High-quality, clearly structured, authoritative content works for both. A well-executed GEO strategy will also improve our SEO performance. Think of GEO as the upgrade layer — we build on our SEO foundation and extend it for the AI era.

How Should We Prioritise?

SEO still drives the majority of organic web traffic and should not be abandoned. GEO is the highest-priority new investment because it intercepts customers at the research and consideration phase — before traditional search even begins. The two strategies reinforce each other.

GEO — Highest New Investment SEO — Maintain & Build
03

How AI Engines Find & Cite Content

To optimise for AI engines, we need to understand how they actually work. The process is fundamentally different from traditional search — and understanding it explains why the GEO tactics we use are the ones we use.

🔍
Query Received
User asks a natural-language question in ChatGPT, CoPilot, or AI Overview
🌐
Web Crawled
AI retrieves and reads multiple web pages it deems relevant and trustworthy
🤖
Answer Synthesised
AI generates a new answer blending sources; selects which to cite
📎
Citation Shown
User sees the answer with clickable source links — our citation opportunity

What Makes AI Engines Choose to Cite a Source?

AI engines don't rank pages the way Google does. They evaluate content for trustworthiness, clarity, and direct relevance to the query. The selection criteria, while proprietary, are well-understood through research:

📋 Content Signals

  • Direct, concise answer to the question asked
  • Question phrased as a heading (conversational)
  • Short, clear paragraphs (not dense walls of text)
  • Lists and structured formats (easy to parse)
  • Specific figures, dates, and named entities

🏛 Authority Signals

  • Domain authority (established, trusted website)
  • Author credentials or institutional affiliation
  • Backlinks from other reputable sources
  • Structured data (schema markup) present
  • Content regularly updated with accurate information
🔑
Key Insight for UOB

UOB already has domain authority as an established financial institution. Our opportunity is on the content signals side — restructuring existing pages to be more readable and parseable by AI engines, without rewriting everything from scratch.

04

What Machine-Readable Content Looks Like

Machine-readable content is written for both humans and AI systems simultaneously. It answers questions directly, uses structured formatting, and includes metadata that helps AI engines understand and trust what they're reading.

A Common Content Pattern — Illustrative Example Only

The following is a generic illustration of a content structure that AI engines struggle to extract citable answers from. It is not a reference to any specific UOB page. Dense, promotional copy with no clear Q&A structure is a pattern seen across many industries — and it's one we can improve.

HTML — Before Generic product page (illustrative example — before optimisation)
<!-- ❌ BEFORE: Typical promotional page structure -->

<h1>[Product Name] Loan</h1>

<p>[Product Name] is a versatile business financing solution
designed to help businesses achieve their goals.
With competitive rates and flexible repayment options,
[Product Name] supports your business at every stage.</p>

<h2>Key Features</h2>
<ul>
  <li>Loan amount up to $500,000</li>
  <li>Flexible tenure options</li>
  <li>Fast approval process</li>
</ul>

<!-- Problems:
  - H1 is a product name, not a question or topic
  - No direct answers to "what is", "how much", "how to apply"
  - Vague language ("competitive", "flexible") — AI can't cite this
  - No FAQ structure
  - No schema markup
  - No date/accuracy signals
-->
GEO-optimised version — AI engines can extract and cite this

Questions as headings. Direct answers. Specific numbers. Author and date signals. This is what gets cited.

HTML — After Generic product page (illustrative example — after GEO optimisation)
<!-- ✅ AFTER: GEO-optimised structure -->

<!-- Authority signal: who wrote it, when -->
<meta name="author" content="UOB Business Banking Team">
<meta name="last-updated" content="2026-01-15">

<!-- Conversational H1 — matches how users ask AI -->
<h1>SME Business Loans: UOB BizMoney Explained</h1>

<!-- Direct answer paragraph — first 2–3 sentences are citable -->
<p>UOB BizMoney is an unsecured business loan for SMEs,
offering loans from <strong>S$10,000 to S$500,000</strong> with tenures
of 1 to 5 years. Interest rates start from <strong>effective 9.8% p.a.</strong>
Approval can be granted within 1 business day for eligible businesses.</p>

<!-- FAQ section with question-phrased headings -->
<section>
  <h2>Who is eligible for a UOB BizMoney loan?</h2>
  <p>Businesses with at least 1 year of operating history
  and minimum annual revenue of S$300,000 can apply for
  UOB BizMoney. Both sole proprietors and private limited
  companies are eligible.</p>

  <h2>How do I apply for a UOB business loan?</h2>
  <p>Apply online via the UOB Business Banking portal,
  visit any UOB branch, or contact our Business Banking team.
  Required documents include your business registration
  and the last 6 months of bank statements.</p>
</section>
🔧
What is Schema Markup?

Schema markup is hidden code (invisible to website visitors) that labels our content so AI engines and search engines understand it precisely. Think of it as a translation layer — our page speaks to humans in paragraphs, and speaks to machines in structured data. For the full vocabulary of schema types, visit Schema.org.

JSON-LD — Schema Markup FAQ schema for a UOB product page
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "name": "UOB BizMoney Business Loan — FAQs",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "How much can I borrow with UOB BizMoney?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "UOB BizMoney offers unsecured business loans from S$10,000 to S$500,000 for eligible SMEs, with repayment tenures of 1 to 5 years."
      }
    },
    {
      "@type": "Question",
      "name": "What is the interest rate for UOB business loans?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "UOB BizMoney interest rates start from an effective rate of 9.8% per annum. The final rate depends on your business profile, loan amount, and tenure selected."
      }
    }
  ]
}
</script>

/* This code is invisible to website visitors but is read
   directly by AI engines and Google to understand our FAQs
   and generate accurate cited answers. */
The FAQ Structure Formula

FAQs are the single most effective GEO content format. AI engines are trained to look for question-answer pairs and trust them as authoritative responses. Here is the exact formula to follow.

The 5-Part Formula for a GEO-Ready FAQ

  1. 1
    Question phrased exactly as a user would ask an AI
    ❌ Weak heading: UOB BizMoney Loan Features
    ✅ GEO heading: How much can I borrow with a UOB business loan?
  2. 2
    Answer the question directly in the first sentence
    ✅ Good: UOB BizMoney offers loans from S$10,000 to S$500,000 for eligible SMEs.
    ❌ Bad: At UOB, we believe in empowering businesses to grow...
  3. 3
    Include specific, citable facts (numbers, dates, names)
    Vague language ("competitive rates", "flexible terms") cannot be cited by AI. Specific facts ("effective 9.8% p.a.", "approval in 1 business day") can be cited directly.
  4. 4
    Keep answers to 40–80 words for simple facts
    AI engines prefer concise, dense answers. Long paragraphs dilute the signal. Use short answer + link to full details if needed.
  5. 5
    Add schema markup to reinforce the structure
    FAQ schema (shown in the previous tab) tells AI engines this is a formally structured Q&A, increasing citation priority significantly.
05

Research & Ideation for AI-Approved Articles

Writing for GEO starts well before the first word is typed. The research phase determines whether our content will be cited — because we need to understand exactly how our target audience is phrasing questions to AI engines, and then write content that precisely answers those questions.

  1. 1
    Map Our Topics to Real AI Queries

    Go to ChatGPT or Google and type queries our customers would actually use. Note the exact phrasing — not "business loan" but "what's the easiest way to get an SME loan". These conversational phrasings become our FAQ headings.

    Example — UOB GWB Query Research Instead of writing about "Corporate Account Opening", research what people actually ask: "How do I open a business bank account quickly?" · "Which bank is easiest for company account opening?" · "What documents do I need to open a corporate account?"
  2. 2
    Use BrightEdge to Validate Search Demand & Measure AI Visibility

    BrightEdge allows us to identify the actual search demand for specific queries mapped to our pages — so we prioritise topics with proven volume. Beyond keyword data, BrightEdge also tracks our GEO performance metrics: mention rate (how often UOB appears in AI responses) and citation rate (how often UOB appears with a clickable link). This data tells us which pages are already earning AI visibility and where the gaps are.

    📊
    Mention Rate

    % of AI responses where UOB is named in any form. Tracks brand presence in the AI research phase.

    🔗
    Citation Rate

    % of AI responses where UOB appears with a clickable link. A stronger signal — drives traffic and authority.

  3. 3
    Audit What AI Currently Says About UOB

    Run each priority query through 2–3 AI engines. Ask: Is UOB cited? If yes, what content is being cited — and is it accurate? If not, what is being cited instead? This tells us both the content gap and the accuracy risk.

    UOB Is Cited

    Verify accuracy. Note which page is being cited. Plan to strengthen that page further.

    ⚠️
    UOB Is Not Cited

    Identify the gap. Find or create a UOB page that directly answers this query. Optimise it for GEO.

  4. 4
    Identify the E-E-A-T Angle for Our Content

    E-E-A-T stands for Experience, Expertise, Authoritativeness, Trustworthiness — Google and AI engines both weight these signals heavily when choosing what to cite. For each piece of content, ask: how does this demonstrate UOB's genuine expertise? See Appendix A2 for a full breakdown of each signal.

    For PFS (Retail Banking)

    • Cite actual UOB product rates and terms
    • Reference local regulatory bodies where relevant
    • Compare options objectively (builds trust)
    • Include last-updated date on all pages

    For GWB (Business Banking)

    • Name specific UOB schemes (e.g. ESG financing)
    • Reference government partnership programmes
    • Provide step-by-step process details
    • Include eligibility criteria precisely
  5. 5
    Plan the Content Structure Before Writing

    Use this template for every new GEO article or FAQ page:

    Content Brief Template
    // GEO Content Brief — fill this before writing
    
    Target query:   "How do SMEs apply for a working capital loan?"
    Product:         UOB BizMoney / SME Working Capital Loan
    User intent:     Research (comparing options before deciding)
    Direct answer:   "SMEs can apply via UOB website, branch, or phone..."
    
    H1: "SME Working Capital Loans: How to Apply with UOB"
    H2s (FAQ format):
      "What is a working capital loan and do I need one?"
      "How much can SMEs borrow for working capital?"
      "What are the eligibility requirements for UOB's SME loan?"
      "How long does UOB take to approve a working capital loan?"
      "What documents do I need to apply?"
    
    Key facts to include: loan range, rate, tenure, eligibility, timeline
    Schema type:     FAQPage + FinancialProduct
    Internal links:  BizMoney product page, branch locator, application form
    Last updated:    Always include publish + review date
  6. 6
    Test & Iterate After Publishing

    After publishing, wait 4–6 weeks then run the same queries in AI engines again. Did citation behaviour change? Which engine cited us? What did they extract? Track this in BrightEdge to measure mention rate and citation rate movement. GEO is a continuous cycle, not a one-time task.

06

Enhancing Existing Product Pages for GEO

We don't need to build new pages from scratch. Most UOB product pages already have the foundational content — they just need to be restructured and enhanced for AI readability. Here is the exact playbook.

💡
The 20% Rule

In most cases, 20% of changes to a page will produce 80% of the GEO improvement. The most impactful changes are: restructuring headings to questions, adding a FAQ section, and adding schema markup. Start there before anything else.

The GEO Page Enhancement Audit — 8 Checks

Run every existing product page through these 8 checks. Each "No" is an optimisation opportunity.

# Audit Check What to Look For Priority
1 Conversational H1 Does the main heading sound like a question users ask AI? Critical
2 Direct Answer Paragraph Do the first 2–3 sentences directly answer the core question? Critical
3 FAQ Section Present Is there a Q&A block with question-phrased H2 headings? Critical
4 Schema Markup Is FAQPage or FinancialProduct schema present in page code? Critical
5 Specific Facts Are exact rates, amounts, and eligibility criteria stated? High
6 Last Updated Date Is there a visible "last updated" or "as of" date? High
7 Internal Linking Does it link to related UOB pages (application, calculator, etc.)? Medium
8 Mobile & Speed Does the page load quickly and render well on mobile? Medium
Real UOB Example — eGuarantee@Gov Page Enhancement The UOB eGuarantee@Gov page was restructured to lead with "What is eGuarantee@Gov and how does UOB support it?" as the H1, followed by a direct paragraph defining the government scheme and UOB's role. An FAQ section was added covering "Who can use UOB eGuarantee@Gov?", "What are the guarantee limits?" and "How long does approval take?" — all phrased as user questions. FAQPage schema was added to the page code. As a result, UOB began appearing in AI citations for government guarantee-related business banking queries.
07

All GEO Elements — Defined & Illustrated

Every GEO-optimised page is built from specific elements. Below is a complete reference of every element — what it is, why it matters, and what it looks like in practice. Click any element to expand its full detail.

● Required ● Recommended ● Optional but Valuable
● Required
FAQ Structure with Question Headings
The most powerful single GEO element. Questions phrased exactly as a user would ask an AI become AI-citable headings.

What It Is

A section of your page where each heading is a natural-language question, followed by a short, direct answer. The question phrasing must match how real users talk to AI — conversational, specific, complete sentences.

Why It Works

AI engines are fundamentally question-answering systems. When they see a page structured as Q&A pairs, they recognise it as a reliable answer source and are significantly more likely to extract and cite from it.

What It Looks Like

HTML
<h2>What is the minimum deposit for a UOB fixed deposit?</h2>
<p>The minimum deposit for a UOB SGD fixed deposit is <strong>S$10,000</strong>,
with tenures ranging from 1 month to 36 months. Rates are fixed
for the entire tenure and guaranteed by UOB.</p>
● Required
Structured Data (Schema Markup)
Hidden code that labels our content so AI engines understand it with machine precision. Think of it as a translation layer between our page and the AI.

What It Is

JSON-LD code embedded in our page's <head> or body that uses Schema.org vocabulary to formally declare what our content is — an FAQ, a financial product, an article, an organisation, etc.

Types Relevant to UOB

  • FAQPage — for all Q&A sections
  • FinancialProduct — for loan/card/savings pages
  • HowTo — for application process pages
  • Organization — for UOB's main brand pages
  • Article — for FinLit and blog content

What It Looks Like

JSON-LD
{
  "@type": "FinancialProduct",
  "name": "UOB One Account",
  "provider": { "@type": "Bank", "name": "UOB" },
  "description": "UOB One Account offers up to 7.8% p.a. bonus interest when you credit your salary and spend on a UOB card.",
  "offers": {
    "@type": "Offer",
    "eligibleRegion": "[Market]"
  }
}
● Required
Direct Answer Opening Paragraph
The first 2–3 sentences must directly answer the page's core topic. AI engines often extract just the opening paragraph for their response.

What It Is

A concise, fact-dense introductory paragraph placed immediately after the H1 that answers the page's core question without preamble or fluff.

The Rule

If we removed everything after the first paragraph, would a reader still get a complete, accurate, citable answer? If yes — it's a good GEO opening. If no — rewrite it.

Example

✅ GEO-Ready Opening "UOB's SME Working Capital Loan provides businesses with up to S$300,000 in unsecured financing, with loan tenures of 1 to 5 years. Businesses with at least 2 years of operating history and annual revenue above S$500,000 can apply online or at any UOB branch."
❌ Not Citable "At UOB, we are committed to supporting SMEs with a range of comprehensive financing solutions tailored to unique business needs and growth aspirations."
● Required
E-E-A-T Authority Signals
Experience, Expertise, Authoritativeness, Trustworthiness — the four signals AI engines use to decide if a source is credible enough to cite. See Appendix A2 for a full breakdown.

What It Is

A set of content and structural signals that establish UOB as a trustworthy, expert source. These signals are evaluated by both Google and AI engines before content is cited.

🔬 Experience

Demonstrates that the content reflects real, first-hand knowledge. For UOB, this means referencing actual product data, real customer scenarios, and lived operational context — not generic descriptions. Example: citing how many customers use a product, or quoting a real application timeline rather than "fast approval".

🎓 Expertise

Shows subject-matter depth. UOB pages should reflect genuine financial knowledge — precise rates, regulatory nuances, eligibility criteria. Content authored or reviewed by named teams (e.g. "UOB Business Banking Team") signals expertise more strongly than anonymous copy.

🏛 Authoritativeness

Signals that UOB is a recognised authority on the topic — not just an opinion, but an institutional source. This is built through references to regulatory bodies, government schemes, industry associations, and inbound links from credible third-party sources.

🛡 Trustworthiness

The most weighted of the four. Trust signals include: a visible last-updated date, clear disclaimers where relevant, transparent T&C references, accurate rates (not outdated), and a secure, well-structured website. AI engines are trained to deprioritise content that could mislead users.

On-Page Trust Signals

  • "Last updated: [date]" on all pages
  • Author attribution (e.g. "UOB Business Banking Team")
  • References to regulatory bodies
  • Transparent about limitations and T&Cs

Off-Page Authority Signals

  • Links from reputable news and finance sites
  • Mentions on government portals
  • UOB brand recognition (AI engines know UOB)
  • Press releases picked up by credible media
Conversational Heading Structure (H1–H3)
Headings restructured to mirror natural-language queries. The hierarchy of H1 → H2 → H3 maps to topic → subtopic → detail.

The Heading Hierarchy

  • H1
    One per page. States the main topic as users would search it. "SME Business Loans — UOB BizMoney"
  • H2
    Major subtopics. Phrased as questions. "How much can I borrow?" "Who is eligible?"
  • H3
    Specific details. Can be informational. "Required Documents" "Application Steps"

Transformation Example

Before: H2: "Loan Features" → H2: "Eligibility" → H2: "Application Process"
After: H2: "How much can I borrow with UOB BizMoney?" → H2: "Am I eligible for a UOB business loan?" → H2: "How do I apply for a UOB business loan?"
Specific, Citable Facts & Figures
Exact numbers, dates, rates, and named entities. Vague language is invisible to AI engines. Specific facts are what get extracted and quoted.

Why Specificity Matters

When an AI generates an answer, it selects content that can be stated with confidence. "Competitive rates" tells an AI nothing it can confidently cite. "Effective rate from 9.8% p.a." is a fact it can use.

What to Include

  • Exact interest rates (with effective rate p.a.)
  • Loan/deposit amount ranges (min and max)
  • Processing / approval timelines
  • Eligibility criteria (revenue, years in business)
  • Named government schemes and partners
❌ Vague: "Enjoy competitive interest rates on our fixed deposits."
✅ Citable: "UOB SGD Fixed Deposit rates start from 2.85% p.a. for a 12-month tenure, with a minimum deposit of S$10,000."
Strategic Internal Linking
Links between UOB pages that help AI engines understand the relationship between topics and signal the depth of UOB's content ecosystem.

What It Is

Hyperlinks from one UOB page to related UOB pages using descriptive anchor text. This helps AI engines understand that UOB has comprehensive, interconnected content on a topic — a signal of authority.

GEO Internal Linking Rules

  • Link from articles to product pages with specific anchor text
  • Use descriptive anchors: "UOB BizMoney loan" not "click here"
  • Link to the most relevant, GEO-optimised page
  • Include links to application forms and calculators
HTML
<!-- ❌ Bad anchor text -->
To apply, <a href="/apply">click here</a>.

<!-- ✅ Descriptive anchor text -->
Apply online via the
<a href="/business/loans/bizmoney">UOB BizMoney application portal</a>.
GEO-Optimised Meta Description
The page meta description is sometimes used by AI engines as the basis for a citation snippet. It should be a single complete, factual sentence.

What It Is

A 150–160 character summary in the page's HTML <head> that describes the page content. Traditionally used for Google snippets — now also read by AI engines.

GEO Meta Description Formula

[Product/Topic] + [Key Fact/Differentiator] + [Action/Audience] + [Brand]

HTML
<!-- ❌ Generic -->
<meta name="description" content="UOB business banking solutions for SMEs.">

<!-- ✅ GEO-optimised -->
<meta name="description"
  content="UOB BizMoney offers SME loans from S$10,000–S$500,000 at effective rates from 9.8% p.a. Apply online in minutes. For eligible SMEs only.">
Content Freshness Signals
Dates and update markers that tell AI engines our content is current and reliable. Outdated content is deprioritised — especially for financial rates.

Why This Matters for Banking

Interest rates, eligibility criteria, and government schemes change frequently. AI engines are trained to prefer recently updated sources for financial information — because citing a page with an outdated rate would mislead users.

What to Add

  • Visible "Last updated: [Month Year]" near the top of every page
  • dateModified in our page schema markup
  • "Rates accurate as of [date]" disclaimer on rate pages
  • Regular content reviews scheduled (quarterly minimum)
How-To / Step-by-Step Content
Numbered process guides for actions like "how to apply", "how to open", "how to switch". AI engines frequently cite step-by-step formats for procedural queries.

When to Use It

Any query that begins with "How do I…", "How to…", or "Steps to…" is a strong candidate for HowTo content. This format is especially powerful for application and onboarding journeys.

JSON-LD — HowTo Schema
{
  "@type": "HowTo",
  "name": "How to open a UOB business account",
  "step": [
    {
      "@type": "HowToStep",
      "name": "Prepare your documents",
      "text": "Gather your business registration certificate, identification documents of all directors, and company constitution."
    },
    {
      "@type": "HowToStep",
      "name": "Apply online or at a branch",
      "text": "Submit your application via UOB BizSmart or visit any UOB Business Banking Centre."
    }
  ]
}
● Optional but Valuable
Comparison & "Best Of" Content
Content that compares UOB products to alternatives or explains when UOB is the right choice. AI engines cite comparison content frequently for high-intent research queries.

Why Comparison Content Gets Cited

When a user asks AI "what's the best SME loan?", the AI generates a comparison. It will cite pages that have already done that comparison work. If UOB has a page that objectively discusses SME loan options — written with transparency and facts — it becomes citable for exactly these high-intent queries.

The Trust Trade-Off

Comparison content that acknowledges alternative options alongside UOB's strengths is more trusted by AI engines than one-sided promotional content. Intellectual honesty signals authority.

Example Queries Comparison Content Can Win "What is the best business account for an SME?" · "Which business loan offers the lowest rate?" · "Which bank is best for a new startup?"
● Optional but Valuable
Entity Coverage & Named References
Explicitly naming relevant entities — government schemes, regulations, partner programmes, industry bodies — helps AI engines understand UOB's ecosystem and cite it in context.

What This Means

AI engines use "entity graphs" — a map of how topics, organisations, and concepts relate to each other. When UOB content explicitly references named entities (regulatory bodies, government schemes, industry programmes), the AI understands UOB is contextually relevant to those topics and may cite UOB when those entities are queried.

Entities to Reference in UOB Content

Local regulatory bodies Government business schemes SME support programmes ESG financing frameworks Digital banking platforms Industry associations
08

GEO Quick-Start Checklist

Use this checklist for every new piece of content and every existing page being optimised. If all 12 items are ticked, the page is GEO-ready.

The GEO-Ready Page Checklist

🏗 Structure

  • H1 is a natural-language topic or question — phrased how a user would ask an AI, not a product name
  • First paragraph directly answers the core topic — specific, factual, citable in 2–3 sentences
  • FAQ section present with question-phrased H2s — minimum 4–6 questions covering what, who, how, how much, how long
  • Headings follow a logical H1 → H2 → H3 hierarchy — no skipping levels, no duplicate H1s

🔧 Technical

  • FAQPage schema markup added — all Q&A pairs reflected in JSON-LD code in page head
  • FinancialProduct or Article schema added — depending on page type
  • Meta description is factual and specific — includes product name, key stat, and audience (150–160 chars)
  • Page loads fast and is mobile-friendly — Core Web Vitals pass

✍️ Content Quality

  • Specific facts included — exact rates, loan amounts, timelines, eligibility criteria
  • Last updated date visible — "Last updated: [Month Year]" displayed on page and in schema
  • Internal links use descriptive anchor text — to related product pages, applications, calculators
  • No vague or purely promotional language — every paragraph can be paraphrased as a standalone fact
🚀
We're Ready to GEO

A page meeting all 12 criteria is significantly more likely to appear in AI-generated responses than pages that haven't been optimised. This isn't theory — these are the factors that drive citation in real AI engine testing. Apply this consistently across UOB's priority pages and our AI search visibility will grow measurably over the next 2–3 quarters.

A1

Appendix — AEO Explained

Answer Engine Optimisation (AEO) is a third discipline alongside SEO and GEO. While it shares foundational principles with GEO, it targets a different surface area and user behaviour.

Emerging
AEO
Answer Engine Optimisation
🎯Goal: Appear in voice search & featured snippet answers
👤User intent: "Give me a quick, direct answer"
📊Metric: Featured snippet ownership, voice answer rate
🛠Tactics: Concise Q&A, definition-led content, schema
📈Attribution: Limited — often zero-click, no visit generated
Speed: 4–8 weeks for featured snippet changes
Key Difference vs GEO
AEO vs GEO
Where they diverge
📍Surface: AEO targets Google featured snippets and voice assistants. GEO targets conversational AI engines (ChatGPT, AI Overviews, CoPilot).
📝Content format: AEO favours single-sentence definitions. GEO favours multi-paragraph Q&A with contextual depth.
🔗Traffic: AEO often generates zero-click impressions (answer shown, no visit). GEO citations include clickable links that drive web visits.
🤝Overlap: Both benefit from FAQ structure, schema markup, and direct answers. GEO content inherits AEO benefits automatically.
💡
Our Recommended Approach

We do not need a separate AEO strategy. A well-executed GEO content strategy — FAQ structure, direct answers, schema markup — will automatically capture AEO benefits. AEO should be thought of as a dividend of good GEO practice, not a separate workstream.

When AEO Becomes Relevant on Its Own

AEO becomes a distinct focus when we want to own specific definitional queries — such as "What is a working capital loan?" or "What is ESG financing?" — where the answer is a single crisp definition rather than a multi-faceted FAQ. These are typically brand-building, top-of-funnel queries where the goal is awareness, not immediate conversion.

GEO — Primary Focus SEO — Maintain & Build AEO — Inherited Benefit
A2

Appendix — E-E-A-T Deep Dive

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness. Originally a Google Search Quality framework, it is now equally relevant to how AI engines evaluate content credibility before deciding to cite it. Understanding each signal in depth is essential for any team producing content for UOB's digital properties.

📌
Why E-E-A-T Matters for AI Citation

AI engines are trained to avoid citing content that could mislead users — especially in high-stakes domains like financial services, health, and law (collectively known as "YMYL" — Your Money or Your Life topics). Banking content falls squarely in this category. This means UOB pages are held to a higher credibility bar than general content, and E-E-A-T signals are what allow us to clear that bar.

  1. E
    Experience — First-Hand Knowledge

    Experience signals show that content reflects genuine, first-hand interaction with the subject — not just theoretical knowledge. For UOB, this means grounding content in real operational data, real product terms, and realistic customer scenarios.

    How to Signal Experience

    • Use real product data — actual rates, real approval timelines, genuine eligibility thresholds
    • Reference realistic scenarios — "A business with 2 years of operating history and S$500K revenue would typically..." rather than vague generalisations
    • Include process specifics — exact documents required, number of steps, expected turnaround
    • Avoid generic copy — phrases like "competitive rates" and "flexible options" signal absence of experience; specific numbers signal its presence
    ❌ Low Experience Signal: "UOB offers competitive rates and flexible repayment options for SMEs."

    ✅ High Experience Signal: "UOB BizMoney offers unsecured loans from S$10,000 to S$500,000 at effective rates from 9.8% p.a., with approval typically within 1 business day for eligible applicants."
  2. E
    Expertise — Subject-Matter Depth

    Expertise signals demonstrate that the content was produced by or in consultation with someone who genuinely knows the subject. For financial content, this is about showing institutional knowledge — not just marketing copy.

    How to Signal Expertise

    • Author attribution — label content as "by the UOB Business Banking Team" or reviewed by a named specialist
    • Regulatory nuance — reference applicable regulatory frameworks, compliance requirements, or licensing conditions where relevant
    • Technical depth — explain how a product actually works, not just its benefits. What is the calculation method for interest? What triggers early repayment fees?
    • Acknowledge complexity — expert content doesn't oversimplify. Saying "rates vary based on your business profile, loan amount, and tenure" is more expert than "great rates available"
  3. A
    Authoritativeness — Institutional Recognition

    Authoritativeness is about whether the wider web and ecosystem recognise UOB as a credible source on a topic. It is the most externally-driven signal — built over time through presence, mentions, and inbound references.

    How to Signal Authoritativeness

    • References to authoritative bodies — citing local central banks, financial regulators, government business agencies, and industry associations demonstrates that UOB operates within a recognised institutional framework
    • Named government schemes — being identified as a participating institution in official government programmes (e.g. SME financing schemes, green loan frameworks) is a strong authority signal
    • Third-party mentions — press coverage, financial news features, and links from credible non-UOB sources all build external authority
    • Awards and recognition — industry awards, rankings, and certifications cited on-page contribute to authority perception
  4. T
    Trustworthiness — The Most Weighted Signal

    Google's own guidelines state that Trustworthiness is the most important of the four E-E-A-T dimensions — and this is especially true for financial content. AI engines inherit this weighting. A page that is inaccurate, misleading, or difficult to verify will not be cited, regardless of how well-written it is.

    How to Signal Trustworthiness

    • Last updated date — a visible, accurate "Last updated: [Month Year]" is one of the single highest-impact trust signals for financial content
    • Rate accuracy disclosures — "Rates accurate as of [date]. Subject to change without notice." tells AI engines the information is time-stamped and maintained
    • Transparent T&Cs — linking to full terms and conditions, and acknowledging conditions and limitations (e.g. "for eligible businesses only") increases trust rather than diminishing it
    • No misleading claims — superlatives without evidence ("best rates in the market") actively harm trust scores. Specific, verifiable claims are always preferable
    • Secure, accessible website — HTTPS, fast load times, and mobile-friendliness are baseline trust signals at the technical layer
    • Schema markup — structured data tells AI engines that the content has been formally labelled and is intended to be machine-readable, which itself is a trust signal
    ⚠️
    The Risk of Outdated Financial Content

    If an AI engine cites a UOB page showing an outdated interest rate or discontinued product, it damages both the customer experience and UOB's credibility. Regular content audits — at minimum quarterly for rate-sensitive pages — are not a nice-to-have; they are a trust maintenance requirement.

E-E-A-T Is a Team Effort

No single team owns E-E-A-T. Experience and Expertise come from product and content teams. Authoritativeness is built by comms, PR, and partnerships. Trustworthiness is maintained by compliance, legal, and web operations. Coordinating these signals across teams is what separates a credible digital presence from one that AI engines overlook.